European Multicenter Study on Analytical Performance of Veris HIV-1 Assay

ABSTRACT

The analytical performance of the Veris HIV-1 assay for use on the new, fully automated Beckman Coulter DxN Veris molecular diagnostics system was evaluated at 10 European virology laboratories. The precision, analytical sensitivity, performance with negative samples, linearity, and performance with HIV-1 groups/subtypes were evaluated. The precision for the 1-ml assay showed a standard deviation (SD) of 0.14 log₁₀ copies/ml or less and a coefficient of variation (CV) of ≤6.1% for each level tested. The 0.175-ml assay showed an SD of 0.17 log₁₀ copies/ml or less and a CV of ≤5.2% for each level tested. The analytical sensitivities determined by probit analysis were 19.3 copies/ml for the 1-ml assay and 126 copies/ml for the 0.175-ml assay. The performance with 1,357 negative samples demonstrated 99.2% with not detected results. Linearity using patient samples was shown from 1.54 to 6.93 log₁₀ copies/ml. The assay performed well, detecting and showing linearity with all HIV-1 genotypes tested. The Veris HIV-1 assay demonstrated analytical performance comparable to that of currently marketed HIV-1 assays. (DxN Veris products are Conformité Européenne [CE]-marked in vitro diagnostic products. The DxN Veris product line has not been submitted to the U.S. FDA and is not available in the U.S. market. The DxN Veris molecular diagnostics system is also known as the Veris MDx molecular diagnostics system and the Veris MDx system.)

INTRODUCTION

Monitoring of viral load (VL) levels is an important part of treatment for patients living with human immunodeficiency virus (HIV). With HIV estimated to affect 37 million people worldwide, the number of patients receiving antiretroviral therapy (ART) and therefore requiring VL monitoring topped 17.0 million in 2015 (1). Current guidelines (including those from the World Health Organization [WHO], the Centers for Disease Control and Prevention [CDC], the U.S. Department of Health and Human Services [USDHHS], and the European AIDS Clinical Society [EACS]) recommend starting all HIV-infected people on ART regardless of VL status or immunologic CD4 level (2–5). The goal is to continually expand the numbers of people able to receive treatment by shifting more money to providing ART. Earlier treatment of patients combined with highly active antiretroviral therapy (HAART) has shown improved suppression of HIV viral loads over time and improved long-term survival (2–5). VL testing is recommended not only for monitoring antiviral suppression of VL but also as the preferred method for detection/confirmation of virologic failure, prediction of clinical progression, and identification of treatment adherence problems (3). Additionally, some antiretroviral medications require pretreatment viral load levels of <100,000 copies/ml in order to be used (3). For patients on treatment, “not detected” or “detected, not quantified” results are required, indicating both compliance with therapy and efficacy of the current regimen. Accurate, sensitive, reliable, and automated easy-to-use viral load monitoring is essential for patient management.

Beckman Coulter's novel DxN Veris molecular diagnostics system (DxN Veris system) is a fully automated system for the quantitative analysis of molecular targets. The system integrates sample introduction, nucleic acid extraction, reaction setup, real-time PCR amplification and detection using TaqMan chemistry, and results interpretation. The system incorporates true random access, allowing flexibility in the number of tests and number of assays that can be run at the same time. The Veris HIV-1 assay is an RNA- and quantitative nucleic acid amplification-based assay for HIV-1, calibrated to the 3rd WHO International Standard for human immunodeficiency virus type 1 (NIBSC 10/152), for use on the DxN Veris system.

The objective of this study was to evaluate the analytical performance of the Veris HIV-1 assay on the DxN Veris system at multiple sites in the European Union.

RESULTS

Precision.

Table 1 shows the SD and percent coefficient of variation (CV) for each component of the reproducibility evaluation for the combined data from 9 sites performing testing for the 1-ml Veris HIV-1 assay. The reproducibility evaluation expected a total of 40 measurements for each panel from each site, for a total of 360 measurements. Panel 1 results are not included in the table; all results returned a value of “not detected” (ND) (n = 320, 100% ND [1 site did not test panel 1]). Results for panel members 2 to 5 showed an SD of 0.14 log₁₀ copies/ml or less for each level tested and a CV decreasing from 6.1% at the low end to 0.7% at the high end. The majority of the imprecision was from within-run and between-site variability. Mean and SD values are given in log₁₀ copies/ml.

TABLE 1.

Precision of the 1-ml and 0.175-ml Veris HIV-1 assays

Assay	Panel	Result summary		SD (log10 copies/ml)a						Total
				Within run	Between run	Within day	Between day	Between ARP lot	Between site
		n	Mean log10 copies/ml							SD (log10 copies/ml)	%CV
1 ml	2	360	2.3	0.12	0.00	0	0.01	0.04	0.07	0.14	6.1
	3	360	3.0	0.07	0	0	0.00	0.02	0.06	0.09	3.0
	4	358	5.2	0.04	0	0	0.01	0.03	0.04	0.06	1.2
	5	360	6.9	0.03	0	0	0.00	0.02	0.03	0.05	0.7
0.175 ml	1	80	3.3	0.13	0	0	0.05	0.03	0.09	0.17	5.2
	2	80	5.1	0.13	0.02	0.02	0.06	0	0.03	0.15	2.9
	3	80	6.8	0.05	0	0	0.04	0.01	0.04	0.08	1.2

^a0 log₁₀ copies/ml, default value when estimated variance was negative.

Table 1 also shows the SD and percent CV for each component of the reproducibility evaluation for the combined data from 2 sites performing testing for the 0.175-ml HIV-1 assay. Results showed an SD of 0.17 log₁₀ copies/ml or less for each level tested and a CV decreasing from 5.2% at the low end to 1.2% at the high end. Mean and SD values are given in log₁₀ copies/ml.

Analytical sensitivity (LOD).

Positive hit rates, with positive defined as quantified (≥35 copies/ml) or detected but not quantified (<35 copies/ml), for the dilution panel for each site completing limit of detection (LOD) testing for the 1-ml HIV-1 assay are presented in Table 2. The calculated LOD by probit analysis for each site and all sites combined is also presented in Table 2. The LOD determined by probit analysis for all sites combined for the 1-ml assay was 19.3 copies/ml (95% confidence interval [CI], 17.6 to 21.5 copies/ml).

TABLE 2.

Analytical sensitivity of the 1-ml Veris HIV-1 assay

Site	Nominal input HIV-1 (copies/ml)	No. of samples tested	No. of samples with detected results	Hit rate (%)	LOD (95% CI) by probit analysis (copies/ml)
Bordeaux	71	36	36	100	19.1 (15.2–26.6)
	35	36	36	100
	21	36	35	97
	7	36	30	83
	0	36	0	0
Toulouse	71	36	36	100	19.0 (15.1–26.6)
	35	36	36	100
	21	36	35	97
	7	36	29	81
	0	36	1	3
London	71	36	36	100	19.8 (15.8–27.6)
	35	36	36	100
	21	36	36	100
	7	36	26	72
	0	36	0	0
Aachen	71	36	36	100	20.0 (16.0–27.7)
	35	36	36	100
	21	36	36	100
	7	36	25	69
	0	35	0	0
Berlin	71	36	36	100	19.1 (15.0–27.2)
	35	36	36	100
	21	36	34	94
	7	36	28	78
	0	36	1	3
Barcelona	71	35	35	100	19.9 (16.1–27.3)
	35	36	36	100
	21	35	34	97
	7	36	26	72
	0	36	0	0
Madrid	71	36	36	100	18.3 (14.2–26.9)
	35	36	36	100
	21	36	36	100
	7	36	33	92
	0	36	0	0
All	71	251	251	100	19.3 (17.6–21.5)
	35	252	252	100
	21	251	246	98
	7	252	197	78
	0	251	2	1

Positive hit rates, with positive defined as quantified (≥235 copies/ml) or detected but not quantified (<235 copies/ml), for the dilution panel for each site completing LOD testing for the 0.175-ml HIV-1 assay are presented in Table 3. The calculated LOD by probit analysis for each site and all sites combined is also presented in Table 3. The LOD determined by probit analysis for sites combined for the 0.175-ml assay was 126 copies/ml (95% CI, 111 to 148 copies/ml).

TABLE 3.

Analytical sensitivity of the 0.175-ml Veris HIV-1 assay

Site	Nominal input HIV-1 (copies/ml)	No. of samples	No. of samples with detected results	Hit rate (%)	LOD (95% CI) by probit analysis (copies/ml)
Sheffield	470	36	36	100	128 (108–162)
	148	36	36	100
	118	36	35	97
	59	36	29	81
	0	36	0	0
Madrid	470	36	36	100	124 (103–158)
	148	36	36	100
	118	36	35	97
	59	36	31	86
	0	36	1	3
All	470	72	72	100	126 (111–148)
	148	72	72	100
	118	72	70	97
	59	72	60	83
	0	72	1	1

Performance with negative samples.

A total of 1,357 specimens were available for negative sample performance. This total included 96 samples tested in duplicate from one site (n = 192) and 1,165 negative pooled samples tested at 6 other sites. The number of samples testing as not detected was 1,346 (99.2%), while 11 (0.8%) samples tested as detected. The 11 samples with detected results were from 4 different sites, and all had values below the LOD (<35 copies/ml) except one, which had a value of 60 copies/ml. The values for the 10 samples with detected results below the LOD were extrapolated as 8.5, 7.7, 7.1, 3.8, 3.7, 3.0, 2.9, 2.7, 2.7, and 2.1 copies/ml.

Linearity and performance with HIV-1 genotypes.

Summary results for the dilution linearity for the 1-ml assay using a high-titer patient sample performed at each of 9 sites are presented in Table 4. One site also determined dilution linearity for the 0.175-ml assay. For the 1-ml assay, the R² was >0.99 at all sites. The maximum degree of nonlinearity was ±0.22 log₁₀ copies/ml at all sites. For the 0.175-ml assay, the R² was >0.99 and the maximum degree of nonlinearity was ±0.13 log₁₀ copies/ml.

TABLE 4.

Dilution linearity with patient samples for Veris HIV-1 assay

Site and assaya	Range evaluated (log10 copies/ml)	n	RSME	R2	Slope (95% CI)	Intercept (95% CI)	Maximum degree of nonlinearity (log10 copies/ml)
Bordeaux	2.79–6.84	18	0.110	0.995	1.05 (1.01–1.09)	−0.22 (−0.41 to −0.03)	±0.116
Toulouse	2.68–6.68	20	0.137	0.993	1.10 (1.06–1.15)	−0.56 (−0.78 to −0.34)	±0.171
Sheffield	1.54–6.93	20	0.086	0.998	0.91 (0.89–0.93)	0.67 (0.57 to 0.76)	—b
Milan Niguarda	2.62–6.62	19	0.122	0.993	1.05 (1.01–1.09)	−0.19 (−0.39 to 0.02)	±0.149
Milan L. Sacco	2.73–6.73	20	0.124	0.994	1.07 (1.03–1.11)	−0.39 (−0.59 to −0.18)	±0.150
Aachen	2.05–6.04	16	0.046	0.999	1.12 (1.10–1.15)	−0.23 (−0.32 to −0.15)	±0.025
Berlin	1.86–6.86	23	0.162	0.991	1.02 (0.98–1.06)	0.07 (−0.13 to 0.27)	±0.213
Barcelona	1.91–6.91	21	0.090	0.997	1.07 (1.04–1.09)	−0.39 (−0.52 to −0.25)	±0.122
Madrid
1 ml	2.89–6.89	20	0.137	0.993	1.15 (1.11–1.20)	−0.93 (−1.16 to −0.70)	±0.111
0.175 ml	2.90–6.90	18	0.129	0.993	1.16 (1.11–1.21)	−1.00 (−1.25 to −0.74)	±0.126

^aAll sites except Madrid tested only the 1-ml assay.

^b—, linear model best, nonlinear models not significantly better at a P level of 0.05.

Summary results for performance with HIV-1 group M (various subtypes and circulating recombinant forms [CRFs]), group N, and group O for the data from one site performing testing are shown in Table 5. The regression plots for each subtype tested are shown in Fig. 1. The maximum degree of nonlinearity was ±0.12 log₁₀ copies/ml for all subtypes tested.

TABLE 5.

Performance with HIV-1 genotypes for the Veris MDx HIV-1 assay

HIV-1 subtype	Range evaluated (log10 copies/ml)	n	RSME	R2	Maximum degree of nonlinearity (log10 copies/ml)
A	1.78–5.01	14	0.090	0.993	±0.068
C	2.11–5.07	15	0.139	0.985	—a
D	1.77–4.81	15	0.109	0.991	±0.078
F	2.09–5.04	15	0.102	0.991	±0.119
H	1.96–5.39	15	0.147	0.985	±0.111
AE	1.93–5.39	14	0.066	0.997	—
AG	1.93–5.44	15	0.071	0.997	—
N	2.00–5.44	15	0.140	0.985	—
O	1.89–5.01	15	0.061	0.997	—

^a—, linear model best, nonlinear models not significantly better at a P level of 0.05.

FIG 1.

Linear regression plots for Beckman Coulter-supplied dilutions of HIV-1 RNA of various group M subtypes, group N, and group O. Data are from one testing site, with plots showing observed result (log₁₀ copies/ml) versus expected result (log₁₀ copies/ml). cp, copies.

DISCUSSION

We evaluated the analytical performance of the Veris HIV-1 assay for use on the DxN Veris system, a new, fully automated system for the quantitation of molecular targets. Performance was evaluated for precision (reproducibility), limit of detection, negative samples, linearity, and HIV-1 groups/subtypes at 10 European clinical virology laboratories.

Precision of the Veris HIV-1 assay showed an SD of ≤0.14 log₁₀ copies/ml or lower at each level tested, with CVs of ≤6.1%. These results match the reported SDs found in the competitor package inserts/instructions for use (IFUs) (6–8) and results found in various independent evaluations of performance claims. In one evaluation of the Abbott RealTime HIV-1 assay, inter- and intra-assay CVs ranged from 1.11 to 2.83% and 1.19 to 4.75%, respectively, for samples with viral loads at 50,000, 5,000, and 500 copies/ml (9). Another evaluation showed both inter- and intra-assay CVs at low VLs to be <10% for the Abbott RealTime HIV-1 assay and <15% for the Siemens Versant HIV assay (10). An evaluation of inter- and intra-assay precision showed total SDs for the Roche Cobas HIV-1 test to be 0.13 log₁₀ copies/ml for both a low (2.3 log₁₀ copies/ml) and a high (4.4 log₁₀ copies/ml) dilution level of the 2nd WHO International Standard for HIV-1 (11).

Analytical sensitivity showed an LOD value based on probit analysis for all sites combined of 19.3 copies/ml for the 1-ml Veris HIV-1 assay (claimed LOD, 35 copies/ml) for the HIV-1 subtype tested. These results show that the Veris HIV-1 assay has a sensitivity comparable to that claimed for the Roche Cobas TaqMan HIV-1 test, v2.0 (claimed LOD, 20 copies/ml) (6), and a sensitivity lower than that claimed for both the Abbott RealTime HIV-1 assay (claimed LOD, 40 copies/ml) (7) and the Siemens Versant HIV-1 RNA 1.5 assay (claimed LOD, 37 copies/ml) (8). The LOD of the Veris HIV-1 assay is consistent with the low LODs required by current guidelines in order to monitor and verify viral suppression while patients are on treatment (2–5).

Testing of 1,357 negative samples resulted in 11 samples with detected results. All samples had values below the LOD of the Veris assay and, in fact, below 10 copies/ml, except for one sample with results quantified at 60 copies/ml. One possible cause of these samples with detected results is preanalytical contamination during sample preparation; however, no contamination or carryover studies were performed during our study. None of the commercially prepared negative quality control (QC) material, tested daily at each site (n = 212 total tests), was detected (data not shown). Carryover studies should be considered in the future. Current HIV-1 quantitative VL assays, including the Veris HIV-1 assay, are not approved for diagnosis of HIV, only for monitoring VL in patients with HIV-1. Testing guidelines (12, 13) indicate that HIV-1 RNA is recommended for confirmation of indeterminate or equivocal results from fourth-generation (HIV antibody and antigen) screening assays in patients suspected of recent infection who have not yet seroconverted. HIV-1 RNA testing is not recommended for broader-use screening for several reasons, including cost and level of false positives. In patients with acute HIV, VL levels are expected to be quite high and some diagnostic guidelines indicate that results with VLs of <1,000 copies/ml should be regarded with caution (13) in these cases.

Results showed that the assay performs well with a dilution series of HIV-1 samples, including those of various group M subtypes, group M CRFs, group N, and group O. All groups and subtypes tested were detected, and linearity was demonstrated across the evaluated range. The ability to detect across a wide range of HIV-1 subtypes is important. HIV-1 exhibits high genetic variability, and HIV mutational changes can result in subtype polymorphisms and the emergence of sequences capable of confounding tests, possibly leading to underquantification (14). This should be taken into consideration if results do not agree with the clinical picture, and retesting on a different system should be considered (15).

HIV VL monitoring is a critical part of management of patients with HIV. These results have demonstrated that the Veris HIV-1 assay is both precise and sensitive at the clinically important low end. These qualities are important in an HIV VL assay in order to accurately determine successful virologic suppression as well as to detect blips and virologic failure (VF), all defined at low-end VLs. Virologic suppression is generally regarded as levels below assay LOD, which can range from 20 up to 40 copies/ml depending on the assay used. Research is mixed on suppression level, with some studies showing that patients with results in the gray area of “detected, not quantified” may be at risk for VF (16, 17), while other studies have indicated that patients with VL levels up to 50 copies/ml do not show increased risk for VF (18, 19). Additionally, the VF definition level is inconsistent among major treatment guidelines, with levels ranging from 50 copies/ml from the EACS (5), to 200 copies/ml from the CDC (3), and even up to 1,000 copies/ml from the WHO (2). More research is needed to refine the levels required for appropriate viral suppression and VF for use across multiple assays. As interassay concordance is poor at the low end, especially when assays with differing LODs are compared (14, 20, 21), it is recommended that VL monitoring be done using the same assay and that system performance characteristics be considered when making treatment decisions (21, 22). If assay changes are made or a new system is introduced, patients may need to have their baseline VL reestablished in order to make appropriate treatment decisions.

The lower-sample-volume 0.175-ml HIV-1 assay was also evaluated at two sites. Precision data showed the total SD to be ≤0.17 log₁₀ copies/ml or lower at each level tested and the CV to be 5.2% at the low end, decreasing to 1.2% at the high end. Analytical sensitivity showed an LOD value based on probit analysis for 2 sites combined of 126 copies/ml for the HIV-1 subtype tested. This compares to the assay claim of 235 copies/ml. The use of low-volume assays has some drawbacks, mainly due to the higher LODs and higher interassay variability (9). These assays are needed, however, when low-volume testing is required, in particular for pediatric patients and certain populations where blood draw volumes are limited.

In conclusion, the analytical data reported here for the Veris HIV-1 assay for use on the DxN Veris system show the assay to be precise, sensitive, and specific, meeting requirements defined in current treatment guidelines. Additionally, a well-performing lower-volume assay is available for use when required.

MATERIALS AND METHODS

The characteristics of the Veris HIV-1 assay are summarized in Table 6. A total of 10 European sites participated in this evaluation. All sites performed testing using the 1-ml HIV-1 assay, while 2 sites also tested the 0.175-ml assay.

TABLE 6.

Summary of DxN Veris system assay characteristics

Assay or sample vol (ml)a	HIV-1 target region	Amplification and detection method	Sample type	Claimed linear range
1	5′ long terminal repeat	Real-time PCR fluorescence detection	Plasma	35b to 1.1 × 107 copies/ml; 1.54 to 7.04 log10 copies/ml
0.175			Plasma	235b to 1.1 × 107 copies/ml; 2.37 to 7.04 log10 copies/ml

^aNot including dead volume.

^bLimit of detection (LOD) = lower limit of quantitation (LLOQ).

Sample panels provided by Beckman Coulter were prepared from the 3rd WHO International Standard for HIV-1 or WHO-traceable HIV-1 material and were shipped on dry ice and stored at −70°C until testing. Samples provided by the sites were leftover patient samples processed per routine and were tested fresh or stored at −70°C until testing. All samples were K₂ EDTA-plasma. Each site received ethics approval prior to the start, either approval for the use of leftover patient samples or a waiver for the use of such samples. All sites followed the sample protocol of the DxN system HIV-1 assay beta study.

Calibration was required for all testing and was performed once prior to the start at each site. Each site used only one lot of assay reagent packs (ARPs) and calibrators during the entire study, but the same lot was not used at all sites. Overall, four lots of ARPs and two lots of calibrators were used. Multiple lots of extraction purification cartridge B (EPB) were used at each site. Quality controls (negative, low, and high for 1-ml assay and negative and high for 0.175-ml assay; from AcroMetrix, ThermoScientific, Fremont, CA) were run daily and required passing results prior to start of testing.

Precision.

Precision was assessed for both the 1-ml and the 0.175-ml Veris HIV-1 assays. Beckman Coulter provided each site with the same HIV-1 standard dilution series panel prepared from WHO-traceable HIV-1 material (HIV-1 group M, subtype B; AcroMetrix, ThermoScientific, Fremont, CA). The series for the 1-ml assay testing contained five K₂ EDTA-plasma samples with the following approximate concentrations of HIV-1 RNA: 0.0 copies/ml, 71 copies/ml (1.9 log₁₀ copies/ml), 1,000 copies/ml (3.0 log₁₀ copies/ml,100,000 copies/ml (5.0 log₁₀ copies/ml), and 5,900,000 copies/ml (6.8 log₁₀ copies/ml). Nine sites performed precision testing for the 1-ml HIV-1 assay. The series for the 0.175-ml assay testing contained three K₂ EDTA-plasma samples with the following approximate concentrations of the HIV-1 RNA virus: 1,000 copies/ml (3.0 log₁₀ copies/ml), 100,000 copies/ml (5.0 log₁₀ copies/ml), and 5,900,000 copies/ml (6.8 log₁₀ copies/ml). Two sites performed precision testing for the 0.175-ml HIV-1 assay. Each site performed testing in one run per day, with each panel member tested in duplicate in each run, for 20 days, for a total of 40 replicates per panel member per site. Study design was based on Clinical and Laboratory Standards Institute (CLSI) guidelines EP05-A3 and EP15-A3 (23, 24). Within-run, between-run, within-day, between-day, between-lot (ARPs and calibrators), between-site, and total imprecision results were evaluated for each level tested by using nested analysis of variance on the log₁₀-transformed HIV-1 RNA values. SD was calculated for all components of variability, and CV was calculated only for total variability.

Analytical sensitivity.

Limit of detection (LOD) was also assessed for both the 1-ml and 0.175-ml assays. Beckman Coulter provided each testing site with the same HIV-1 standard dilution series panel containing five K₂ EDTA-plasma samples with concentrations of the HIV-1 RNA around the LOD. The levels were 0.0, 7.0, 21.0, 35.0, and 71.0 copies/ml for the 1-ml assay and 0.0, 59.0, 118.0, 148.0, and 470.0 copies/ml for the 0.175-ml assay. Both were prepared with the 3rd WHO International Standard for HIV-1, group M, subtype B. Seven sites performed LOD testing for the 1-ml assay, and 2 sites performed LOD testing for the 0.175-ml assay. Each site performed LOD testing in one run per day, with each panel member tested in 12 replicates, for a total of 3 days, for a total of 36 replicates per panel member per site. The study design was based on the CLSI EP17-A2 guideline (25). The concentration at 95% positive results was determined using probit analysis.

Performance with negative samples.

Performance with negative samples was assessed at 7 sites using a combination of Beckman Coulter-provided K₂ EDTA-plasma samples made from pooled patient samples (BioreclamationIVT, New York, NY) or site-provided unique K₂ EDTA-plasma patient samples tested in duplicate. These specimens were negative for HIV-1 RNA by a comparator assay and anti-HIV-1 antibodies. The numbers (percentages) of samples with detected and not detected results were determined. All sites used the 1-ml assay for testing. Actual values for any samples testing “detected, not quantified” (<35 copies/ml) were calculated.

Linearity and performance with HIV-1 genotypes.

To assess linearity, sites selected a high-titer HIV-1 patient sample and performed 4 to 6 serial dilutions to below the LOD with a known negative patient sample or Basematrix (SeraCare, Milford, MA, USA). Each level (original undiluted sample and dilutions to below LOD) was tested in 4 replicates. The study design was based on CLSI EP06-A guidelines (26). The linear regression line, R², root mean square error (RMSE), and maximum difference from linearity were determined. Nine sites performed testing with the 1-ml assay, and one site performed testing with the 0.175-ml assay.

To assess performance with HIV-1 genotypes, Beckman Coulter provided one testing site with an HIV-1 standard dilution series panel containing four or five K₂ EDTA-plasma dilution samples from 10 HIV-1 genotypes (A, C, D, F, G, H, AE, AG, N, and O). HIV-1 genotype material was obtained from SeraCare (Milford, MA, USA). Each sample dilution from each genotype was tested in triplicate. The linear regression line, maximum difference from linearity, R², and RMSE were determined.